Present systems for extracting a wanted signal from a noisy signal include averaging, selective amplification and/or filtering, synchronized detection (e.g., phase lock loop), direct sequence spread spectrum (DSSS) and digital coding.
Averaging reduces noise over n periods, but the wanted signal is not amplified over n periods. Averaging also requires a repetitive signal over n periods, and some type of trigger, and can be problematic at very low signal levels. Selective amplification and/or filtering are frequency dependent and after the setup time do not improve over time in a given bandwidth nor reduce noise within that bandwidth. Selective amplification and/or filtering also have limited noise rejection capabilities. Synchronized detection is limited to a narrow band, is phase locked to the input signal, and is also problematic at very low signal levels. While DSSS spreads the bit energy over a wide frequency spectrum and the recovery of the data dispreads the energy and makes it appear much above the noise floor, DSSS requires the transmitter and the receiver to operate using the same pseudo-noise (PN) sequence. Digital coding, such as Viterbi coding, forward error correction, etc., may increase the signal-to-noise ratio (SNR), but it does so at the expense of reducing throughput data. As a result, there is no satisfactory circuit that enables the detection of a very low signal buried in noise, that enables the regeneration of a repetitive signal, or that increases the signal over noise of the signal in electrical, electronic, telecommunication or wireless applications. The present disclosure discloses a low noise detection system that overcomes these limitations.